🎓 B.Tech CSE (AI & ML) @ Shri Ramdeobaba College of Engineering & Management, Nagpur — CGPA 9.24/10 | Class of 2027
- 🌱 I'm currently deepening my skills in Deep Learning and Agentic AI systems.
- 🔭 I'm working on AI/ML projects that solve real-world problems — from helping millet farmers get fair prices to large-scale grain quality detection.
- 👯 I'm open to collaborating on machine learning, computer vision, and agentic AI projects.
- 🤔 I'm exploring LLM integration, multi-agent orchestration, and ML deployment.
- 💬 Ask me about Python, PyTorch, XGBoost, and building end-to-end ML pipelines.
- 📫 How to reach me: prembaba2823@gmail.com
A farmer uploads a grain photo and gets a predicted quality grade, expected price, and the best nearby mandi to sell at — powered by 4 parallel AI agents (Quality, Price, Market, Decision) in a Planner + Orchestrator pattern. Trained an EfficientNetB0 classifier on a personally collected ~12K image dataset of ragi, bajra, and jowar; price modelling with XGBoost on real AGMARKNET (Govt. of India) data.
Trained EfficientNetV2-S on 141K images from GrainSet (Nature Scientific Data, 2023) for 7-class cereal-grain quality detection — 98.98% macro F1, 99.23% accuracy. Focal Loss, AdamW + cosine-annealing, systematic hyperparameter tuning across 12 configurations.
Complete ML pipeline on 48,842 samples — EDA, feature engineering, and benchmarking of 6 models. XGBoost hit 87.29% accuracy / 0.93 ROC-AUC; deployed live on Hugging Face Spaces with a model comparison dashboard. [Code]
- Languages: Python, Java, SQL
- ML / DL: Scikit-learn, XGBoost, LightGBM, PyTorch, timm, Keras, Pandas, NumPy
- AI Systems: Agentic AI (Planner + Orchestrator, multi-agent orchestration), Computer Vision (EfficientNet), Model Deployment
- Tools & Deployment: Git, GitHub, Streamlit, Plotly, Hugging Face Spaces, SQLite, CUDA
- Core Concepts: Machine Learning, Deep Learning, DSA, OOP, DBMS, Operating Systems, Computer Networks
- Solved 190+ problems on LeetCode (max rating 1642) — Profile
- ⭐ 3-Star SQL on HackerRank
- 📜 Python 101 for Data Science — IBM | Fundamentals of Deep Learning — NVIDIA
